Mars2020_DataAnalysis

IMPORT

Librairies

library(phyloseq)
library(ggplot2)
library(ggsignif)
library(cowplot)
library(tidyverse)
library(reshape2)
library("plotly") # plot 3D
library("microbiome") # for centered log-ratio
library("coda") # Aitchison distance
library("coda.base") # Aitchison distance
library("vegan") # NMDS
library(pheatmap) # for heatmap

Data

# Set path
path <- "~/Projects/IBS_Meta-analysis_16S"
path.plots <- file.path(path, "data/analysis-individual/PLOTS/plots-Mars-EDA")

# Import phyloseq object
physeq.mars <- readRDS(file.path(path, "phyloseq-objects/physeq_mars.rds"))

# Sanity check
physeq.mars

## phyloseq-class experiment-level object
## otu_table()   OTU Table:         [ 1557 taxa and 69 samples ]
## sample_data() Sample Data:       [ 69 samples by 12 sample variables ]
## tax_table()   Taxonomy Table:    [ 1557 taxa by 7 taxonomic ranks ]
## phy_tree()    Phylogenetic Tree: [ 1557 tips and 1555 internal nodes ]
## refseq()      DNAStringSet:      [ 1557 reference sequences ]

Phylogenetic tree was computed with the package phangorn, and the script was run on a cluster. Let’s check we have correctly generated a phylogenetic tree.

# Look at the tree
plot_tree(physeq.mars, color = "Phylum", ladderize="left")

DEMOGRAPHICS

This dataset has several covariates (gender, age, bmi). We will check whether there is the same distribution of these covariates between healthy and IBS patients.

# Number of individuals in each group (keeping just 1 sample per individual)
metadata <- data.frame(sample_data(physeq.mars)) %>%
  select(host_disease, host_bmi, host_age, host_sex, host_ID) %>%
  group_by(host_ID) %>%
  summarise_all(first)
  
metadata %>%
  count(host_disease)

# Age
metadata %>%
  group_by(host_disease) %>%
  summarize(mean_age=mean(host_age), sd_age=sd(host_age))

wilcox.test(metadata[metadata$host_disease == "IBS", ]$host_age,
            metadata[metadata$host_disease == "Healthy", ]$host_age,
            correct=FALSE) # p=0.84

## 
##  Wilcoxon rank sum test
## 
## data:  metadata[metadata$host_disease == "IBS", ]$host_age and metadata[metadata$host_disease == "Healthy", ]$host_age
## W = 155.5, p-value = 0.8429
## alternative hypothesis: true location shift is not equal to 0

# Gender
metadata %>%
  count(host_disease, host_sex)

# Can't use chi^2 test because one of the values < 5
fisher.test(data.frame("Female" = c(8,21),
                      "Male" = c(4,6),
                      row.names = c("Healthy", "IBS"))) # p=0.7

## 
##  Fisher's Exact Test for Count Data
## 
## data:  data.frame(Female = c(8, 21), Male = c(4, 6), row.names = c("Healthy", "IBS"))
## p-value = 0.6927
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.1020796 3.5642776
## sample estimates:
## odds ratio 
##  0.5801387

# BMI
metadata %>%
  group_by(host_disease) %>%
  summarize(mean_bmi=mean(na.omit(host_bmi)), sd_bmi=sd(na.omit(host_bmi)))

wilcox.test(metadata[metadata$host_disease == "IBS",]$host_bmi,
            metadata[metadata$host_disease == "Healthy", ]$host_bmi,
            correct=FALSE) # 0.98

## 
##  Wilcoxon rank sum test
## 
## data:  metadata[metadata$host_disease == "IBS", ]$host_bmi and metadata[metadata$host_disease == "Healthy", ]$host_bmi
## W = 163, p-value = 0.9757
## alternative hypothesis: true location shift is not equal to 0

ABUNDANCES

1. Absolute abundances

# Plot Phylum
plot_bar(physeq.mars, fill = "Phylum") + facet_wrap("host_disease", scales="free_x") +
  theme_cowplot()+
  theme(axis.text.x = element_text(size = 6, angle=90))+
  labs(x = "Samples", y = "Absolute abundance", title = "Mars dataset (2020)")

# ggsave(file.path(path.plots, "absAbundance_phylum.jpg"), width=13, height=5)

# Plot Class
plot_bar(physeq.mars, fill = "Class")+ facet_wrap("host_disease", scales="free_x") +
  theme_cowplot()+
  theme(axis.text.x = element_text(size = 6, angle=90))+
  labs(x = "Samples", y = "Absolute abundance", title = "Mars dataset (2020)")

Sequencing depth characteristics of the Mars dataset:
- minimum of 1140 total count per sample
- median: 3.255610^{4} total count per sample
- maximum of 6.360210^{4} total count per sample

2. Relative abundances

# Agglomerate to phylum & class levels
phylum.table <- physeq.mars %>%
  tax_glom(taxrank = "Phylum") %>%                     # agglomerate at phylum level
  transform_sample_counts(function(x) {x/sum(x)} ) %>% # Transform to rel. abundance
  psmelt()                                             # Melt to long format

class.table <- physeq.mars %>%
  tax_glom(taxrank = "Class") %>%
  transform_sample_counts(function(x) {x/sum(x)} ) %>%
  psmelt()

sample_data(physeq.mars)$host_ID <- as.character(sample_data(physeq.mars)$host_ID)

# Plot relative abundances
ggplot(phylum.table, aes(x = reorder(Sample, Sample, function(x) mean(phylum.table[Sample == x & Phylum == 'Firmicutes', 'Abundance'])),
                         y = Abundance, fill = Phylum))+
  facet_wrap(~ host_disease, scales = "free") + # scales = "free" removes empty lines
  geom_bar(stat = "identity") +
  theme_cowplot()+
  theme(axis.text.x = element_blank())+
  labs(x = "Samples", y = "Relative abundance", title = "Mars dataset (2020)")

# ggsave(file.path(path.plots, "relAbundance_phylum.jpg"), width=10, height=5)

ggplot(class.table, aes(x = reorder(Sample, Sample, function(x) mean(phylum.table[Sample == x & Phylum == 'Firmicutes', 'Abundance'])),
                        y = Abundance, fill = Class))+
  facet_wrap(~ host_disease, scales = "free") + # scales = "free" removes empty lines
  geom_bar(stat = "identity") +
  theme_cowplot()+
  theme(axis.text.x = element_blank(),
        legend.key.size = unit(0.2, 'cm'),
        legend.text = element_text(size=8))+
  labs(x = "Samples", y = "Relative abundance", title = "Mars dataset (2020)")

# ggsave(file.path(path.plots, "relAbundance_class.jpg"), width=12, height=5)

# As curiosity, group samples per host_ID (is there a big difference between collection times?)
ggplot(phylum.table, aes(x = Sample, y = Abundance, fill = Phylum))+
  facet_wrap(~ host_ID, scales = "free_x", nrow=2) + # scales = "free" removes empty lines
  geom_bar(stat = "identity", width=1) +
  theme_cowplot()+
  theme(axis.text.x = element_blank(),
        panel.spacing = unit(0.1, "lines"),
        strip.text = element_text(size=8),
        legend.key.size = unit(0.2, 'cm'),
        legend.text = element_text(size=8))+
  labs(x = "Samples", y = "Relative abundance", title = "Mars dataset (2020)")

3. Firmicutes/Bacteroidota ratio

# Extract abundance of only Bacteroidota and Firmicutes
relevant.covariates <- c('Sample', 'Abundance', 'host_disease', 'Phylum', 'host_ID', 'host_subtype', 'Collection', 'host_sex')

bacter <- phylum.table %>%
  filter(Phylum == "Bacteroidota") %>%
  select(all_of(relevant.covariates)) %>%
  dplyr::rename(Bacteroidota = Abundance) %>%
  select(-Phylum)

firmi <- phylum.table %>%
  filter(Phylum == "Firmicutes") %>%
  select(all_of(relevant.covariates)) %>%
  dplyr::rename(Firmicutes = Abundance) %>%
  select(-Phylum)

# Calculate log2 ratio Firmicutes/Bacteroidota
ratio.FB <- left_join(x=bacter, y=firmi, by=c('Sample', 'host_disease', 'host_ID', 'host_subtype', 'Collection', 'host_sex')) %>%
  relocate(Firmicutes, .after=Bacteroidota) %>%
  # Compute log ratios
  mutate(logRatioFB = log2(Firmicutes/Bacteroidota))


# Plot log2 ratio Firmicutes/Bacteroidota
ggplot(ratio.FB, aes(x = host_disease, y = logRatioFB))+
  geom_violin(aes(fill=host_disease))+
  scale_fill_manual(values=scales::alpha(c("blue", "red"), .3))+
  geom_jitter(width=0.1)+
  geom_signif(comparisons = list(c("Healthy", "IBS")), map_signif_level = TRUE, test="wilcox.test") +
  labs(x = "",  y = 'Log2(Firmicutes/Bacteroidota)')+
  theme_cowplot()+
  theme(legend.position="none")

# ggsave(file.path(path.plots, "ratioFB.jpg"), width=4, height=6)
# Statistical test
wilcox.test(ratio.FB[ratio.FB$host_disease == "IBS","logRatioFB"],
            ratio.FB[ratio.FB$host_disease == "Healthy","logRatioFB"]) # p = 0.6

## 
##  Wilcoxon rank sum exact test
## 
## data:  ratio.FB[ratio.FB$host_disease == "IBS", "logRatioFB"] and ratio.FB[ratio.FB$host_disease == "Healthy", "logRatioFB"]
## W = 579, p-value = 0.63
## alternative hypothesis: true location shift is not equal to 0

# What about per Collection time point?
ggplot(ratio.FB, aes(x = host_disease, y = logRatioFB))+
  geom_violin(aes(fill=host_disease))+
  scale_fill_manual(values=scales::alpha(c("blue", "red"), .3))+
  geom_jitter(width=0.1)+
  facet_wrap(~Collection) +
  geom_signif(comparisons = list(c("Healthy", "IBS")), map_signif_level = TRUE, test="wilcox.test") +
  labs(x = "",  y = 'Log2(Firmicutes/Bacteroidota)')+
  theme_cowplot()+
  theme(legend.position="none")

ggplot(ratio.FB, aes(x = Collection, y = logRatioFB))+
  geom_violin(aes(fill=host_disease))+
  scale_fill_manual(values=scales::alpha(c("blue", "red"), .3))+
  geom_point(size = 2)+
  facet_wrap(~host_disease) +
  geom_line(aes(group=host_ID), lwd=0.2) +
  labs(x = "",  y = 'Log2(Firmicutes/Bacteroidota)')+
  theme_cowplot()+
  theme(legend.position="none")

# ggsave(file.path(path.plots, "ratioFB_paired-data.jpg"), width=4, height=6)

# Plot by IBS subtype
ggplot(ratio.FB, aes(x = host_subtype, y = logRatioFB))+
  facet_wrap(~Collection) +
  geom_violin(aes(fill=host_subtype))+
  scale_fill_manual(values=scales::alpha(c("blue", "brown", "red"), .3))+
  geom_jitter(width=0.1)+
  geom_signif(comparisons = list(c("HC", "IBS-C"), c("HC", "IBS-D")),
              map_signif_level = TRUE, test="wilcox.test", step_increase=0.1) +
  labs(x = "",  y = 'Log2(Firmicutes/Bacteroidota)')+
  theme_cowplot()+
  theme(legend.position="none")

# ggsave(file.path(path.plots, "ratioFB_subtype.jpg"), width=5, height=5)

# Statistical test HC vs IBS-C
wilcox.test(ratio.FB[ratio.FB$Collection == "1st" & ratio.FB$host_subtype == "HC","logRatioFB"],
            ratio.FB[ratio.FB$Collection == "1st" & ratio.FB$host_subtype == "IBS-C","logRatioFB"]) # p = 0.04

## 
##  Wilcoxon rank sum exact test
## 
## data:  ratio.FB[ratio.FB$Collection == "1st" & ratio.FB$host_subtype == "HC", "logRatioFB"] and ratio.FB[ratio.FB$Collection == "1st" & ratio.FB$host_subtype == "IBS-C", "logRatioFB"]
## W = 44, p-value = 0.04069
## alternative hypothesis: true location shift is not equal to 0

wilcox.test(ratio.FB[ratio.FB$Collection == "2nd" & ratio.FB$host_subtype == "HC","logRatioFB"],
            ratio.FB[ratio.FB$Collection == "2nd" & ratio.FB$host_subtype == "IBS-C","logRatioFB"]) # p = 0.9

## 
##  Wilcoxon rank sum exact test
## 
## data:  ratio.FB[ratio.FB$Collection == "2nd" & ratio.FB$host_subtype == "HC", "logRatioFB"] and ratio.FB[ratio.FB$Collection == "2nd" & ratio.FB$host_subtype == "IBS-C", "logRatioFB"]
## W = 56, p-value = 0.917
## alternative hypothesis: true location shift is not equal to 0

# Statistical test HC vs IBS-D
wilcox.test(ratio.FB[ratio.FB$Collection == "1st" & ratio.FB$host_subtype == "HC","logRatioFB"],
            ratio.FB[ratio.FB$Collection == "1st" & ratio.FB$host_subtype == "IBS-D","logRatioFB"]) # p = 0.7

## 
##  Wilcoxon rank sum exact test
## 
## data:  ratio.FB[ratio.FB$Collection == "1st" & ratio.FB$host_subtype == "HC", "logRatioFB"] and ratio.FB[ratio.FB$Collection == "1st" & ratio.FB$host_subtype == "IBS-D", "logRatioFB"]
## W = 60, p-value = 0.7399
## alternative hypothesis: true location shift is not equal to 0

wilcox.test(ratio.FB[ratio.FB$Collection == "2nd" & ratio.FB$host_subtype == "HC","logRatioFB"],
            ratio.FB[ratio.FB$Collection == "2nd" & ratio.FB$host_subtype == "IBS-D","logRatioFB"]) # p = 0.29

## 
##  Wilcoxon rank sum exact test
## 
## data:  ratio.FB[ratio.FB$Collection == "2nd" & ratio.FB$host_subtype == "HC", "logRatioFB"] and ratio.FB[ratio.FB$Collection == "2nd" & ratio.FB$host_subtype == "IBS-D", "logRatioFB"]
## W = 84, p-value = 0.2875
## alternative hypothesis: true location shift is not equal to 0

# Just for curiosty, look by host_sex
ggplot(ratio.FB, aes(x = host_disease, y = logRatioFB))+
  geom_violin(aes(fill=host_disease))+
  scale_fill_manual(values=scales::alpha(c("blue", "red"), .3))+
  geom_jitter(width=0.1)+
  facet_wrap(~host_sex) +
  labs(x = "",  y = 'Log2(Firmicutes/Bacteroidota)')+
  theme_cowplot()+
  theme(legend.position="none")

NORMALIZE DATA

# Sanity check no sample with less than 500 total count
table(sample_sums(physeq.mars)<500) # all FALSE

#____________________________________________________________________
# PHYLOSEQ OBJECT WITH NON-ZERO COMPOSITIONS
physeq.NZcomp <- physeq.mars
otu_table(physeq.NZcomp)[otu_table(physeq.NZcomp) == 0] <- 0.5 # pseudocounts

# Sanity check that 0 values have been replaced
# otu_table(physeq.mars)[1:5,1:5]
# otu_table(physeq.NZcomp)[1:5,1:5]

# transform into compositions
physeq.NZcomp <- transform_sample_counts(physeq.NZcomp, function(x) x / sum(x) )
table(rowSums(otu_table(physeq.NZcomp))) # check if there is any row not summing to 1

# Save object
saveRDS(physeq.NZcomp, file.path(path, "data/analysis-individual/Mars-2020/02_EDA-Mars/physeq_NZcomp.rds"))

#____________________________________________________________________
# PHYLOSEQ OBJECT WITH RELATIVE COUNT (BETWEEN 0 AND 1)
physeq.rel <- physeq.mars
physeq.rel <- transform_sample_counts(physeq.rel, function(x) x / sum(x) ) # divide each count by the total number of counts (per sample)

# check the counts are all relative
# otu_table(physeq.mars)[1:5, 1:5]
# otu_table(physeq.rel)[1:5, 1:5]

# sanity check
table(rowSums(otu_table(physeq.rel))) # check if there is any row not summing to 1

# save the physeq.rel object
saveRDS(physeq.rel, file.path(path, "data/analysis-individual/Mars-2020/02_EDA-Mars/physeq_relative.rds"))

#____________________________________________________________________
# PHYLOSEQ OBJECT WITH COMMON-SCALE NORMALIZATION
physeq.CSN <- physeq.mars
physeq.CSN <- transform_sample_counts(physeq.CSN, function(x) (x*min(sample_sums(physeq.CSN))) / sum(x) )

# sanity check
table(rowSums(otu_table(physeq.CSN))) # check that all rows are summing to the same total

# save the physeq.CSN object
saveRDS(physeq.CSN, file.path(path, "data/analysis-individual/Mars-2020/02_EDA-Mars/physeq_CSN.rds"))


#____________________________________________________________________
# PHYLOSEQ OBJECT WITH CENTERED LOG RATIO COUNT
physeq.clr <- physeq.mars
physeq.clr <- microbiome::transform(physeq.mars, "clr") # the function adds pseudocounts itself

# Compare the otu tables in the original phyloseq object and the new one after CLR transformation
# otu_table(physeq.mars)[1:5, 1:5] # should contain absolute counts
# otu_table(physeq.clr)[1:5, 1:5] # should all be relative

# save the physeq.rel object
saveRDS(physeq.clr, file.path(path, "data/analysis-individual/Mars-2020/02_EDA-Mars/physeq_clr.rds"))

COMPUTE DISTANCES

1. UniFrac, Aitchison, Bray-Curtis and Canberra

First, let’s look at these four distances of interest.

#____________________________________________________________________________________
# Measure distances
getDistances <- function(relPhyseq=physeq.rel, clrPhyseq=physeq.clr, csnPhyseq=physeq.CSN, nzcompPhyseq=physeq.NZcomp){
  # sanity check
  cat("nb samples relPhyseq:", nsamples(relPhyseq), "\n")
  cat("nb samples clrPhyseq:", nsamples(clrPhyseq), "\n")
  cat("nb samples csnPhyseq:", nsamples(csnPhyseq), "\n")
  cat("nb samples nzcompPhyseq:", nsamples(nzcompPhyseq), "\n")
  
  # Compute distances
  print("Unifrac...")
  set.seed(123) # for unifrac, need to set a seed
  glom.UniF <- UniFrac(relPhyseq, weighted=TRUE, normalized=TRUE) # weighted unifrac
  print("Aitchison...")
  glom.ait <- phyloseq::distance(clrPhyseq, method = 'euclidean') # aitchison
  print("Bray des bois...")
  glom.bray <- phyloseq::distance(csnPhyseq, method = "bray") # bray-curtis
  print("Canberra <3...")
  glom.can <- phyloseq::distance(nzcompPhyseq, method = "canberra") # canberra
  
  dist.list <- list("UniF" = glom.UniF, "Ait" = glom.ait, "Canb" = glom.can, "Bray" = glom.bray)
  
  return(dist.list)
}


#____________________________________________________________________________________
# Plot in 2D the distances
plotDistances2D <- function(dlist, ordination="MDS", relPhyseq=physeq.rel, clrPhyseq=physeq.clr, csnPhyseq=physeq.CSN, nzcompPhyseq=physeq.NZcomp){
  plist <- NULL
  plist <- vector("list", 4)
  names(plist) <- c("Weighted Unifrac", "Aitchison", "Bray-Curtis", "Canberra")
  
  print("Unifrac")
  # Weighted UniFrac
  set.seed(123)
  iMDS.UniF <- ordinate(relPhyseq, ordination, distance=dlist$UniF)
  plist[[1]] <- plot_ordination(relPhyseq, iMDS.UniF, color="host_disease")
  
  print("Aitchison")
  # Aitchison
  set.seed(123)
  iMDS.Ait <- ordinate(clrPhyseq, ordination, distance=dlist$Ait)
  plist[[2]] <- plot_ordination(clrPhyseq, iMDS.Ait, color="host_disease")
  
  print("Bray")
  # Bray-Curtis
  set.seed(123)
  iMDS.Bray <- ordinate(csnPhyseq, ordination, distance=dlist$Bray)
  plist[[3]] <- plot_ordination(csnPhyseq, iMDS.Bray, color="host_disease")
  
  print("Canberra")
  # Canberra
  set.seed(123)
  iMDS.Can <- ordinate(nzcompPhyseq, ordination, distance=dlist$Can)
  plist[[4]] <- plot_ordination(nzcompPhyseq, iMDS.Can, color="host_disease")
  
  # Creating a dataframe to plot everything
  plot.df = plyr::ldply(plist, function(x) x$data)
  names(plot.df)[1] <- "distance"
  
  return(plot.df)
}

Now let’s plot!

a) Samples from 1st collection time point (more samples)

#___________________________
# 1ST COLLECTION TIME POINT
#___________________________

# Get the distances & the plot data
dist.mars.1st <- getDistances(relPhyseq = subset_samples(physeq.rel, Collection=="1st"),
                              clrPhyseq = subset_samples(physeq.clr, Collection=="1st"),
                              csnPhyseq = subset_samples(physeq.CSN, Collection=="1st"),
                              nzcompPhyseq = subset_samples(physeq.NZcomp, Collection=="1st"))

## nb samples relPhyseq: 37 
## nb samples clrPhyseq: 37 
## nb samples csnPhyseq: 37 
## nb samples nzcompPhyseq: 37 
## [1] "Unifrac..."
## [1] "Aitchison..."
## [1] "Bray des bois..."
## [1] "Canberra <3..."

plot.df.1st <- plotDistances2D(dlist=dist.mars.1st,
                              relPhyseq = subset_samples(physeq.rel, Collection=="1st"),
                              clrPhyseq = subset_samples(physeq.clr, Collection=="1st"),
                              csnPhyseq = subset_samples(physeq.CSN, Collection=="1st"),
                              nzcompPhyseq = subset_samples(physeq.NZcomp, Collection=="1st"))

## [1] "Unifrac"
## [1] "Aitchison"
## [1] "Bray"
## [1] "Canberra"

# Plot
ggplot(plot.df.1st, aes(Axis.1, Axis.2, color=host_disease))+
  geom_point(size=4, alpha=0.5)  + scale_color_manual(values = c('blue', 'red'))+
  facet_wrap(distance~., scales='free', nrow=1)+
  theme_bw()+
  theme(strip.text.x = element_text(size=20))+
  labs(color="Disease", title="1st collection time point")

# ggsave(file.path(path.plots, "distances4_MDS_1stcollection.jpg"), height = 4, width = 15)

b) All samples

#________________
# ALL DATA
#________________

# Get the distances & the plot data
dist.mars <- getDistances()

## nb samples relPhyseq: 69 
## nb samples clrPhyseq: 69 
## nb samples csnPhyseq: 69 
## nb samples nzcompPhyseq: 69 
## [1] "Unifrac..."
## [1] "Aitchison..."
## [1] "Bray des bois..."
## [1] "Canberra <3..."

plot.df <- plotDistances2D(dlist=dist.mars)

## [1] "Unifrac"
## [1] "Aitchison"
## [1] "Bray"
## [1] "Canberra"

# Plot
p1 <- ggplot(plot.df, aes(Axis.1, Axis.2, color=host_disease))+
  geom_point(size=4, alpha=0.5)  + scale_color_manual(values = c('blue', 'red'))+
  facet_wrap(distance~., scales='free', nrow=1)+
  theme_bw()+
  theme(strip.text.x = element_text(size=20),
        axis.title.x = element_blank())+
  labs(color="Disease")
p2 <- ggplot(plot.df, aes(Axis.1, Axis.2, color=host_subtype))+
  geom_point(size=4, alpha=0.5)  + scale_color_manual(values = c('blue', 'brown', 'red'))+
  facet_wrap(distance~., scales='free', nrow=1)+
  theme_bw()+
  theme(strip.text.x = element_text(size=20),
        axis.title.x = element_blank())+
  labs(color="Disease")
p3 <- ggplot(plot.df, aes(Axis.1, Axis.2, color=Collection))+
  geom_line(aes(group=host_ID), color="black", lwd=0.1)+
  geom_point(size=4, alpha=0.7)  +
  scale_color_manual(values = c('#D95F02', '#7570B3'))+
  facet_wrap(distance~., scales='free', nrow=1)+
  theme_bw()+
  theme(strip.text.x = element_blank())+
  labs(color="Collection ")
ggpubr::ggarrange(p1,p2,p3, nrow=3)

# ggsave(file.path(path.plots, "distances4_MDS_all.jpg"), height = 10, width = 15)

2. Plot in 3D

For better visualization, we will also take a glance at reduction to 3D.

#____________________________________________________________________________________
# Plot 3D ordination
plotDistances3D <- function(d, name_dist){
  
  # Reset parameters
  mds.3D <- NULL
  xyz <- NULL
  fig.3D <- NULL
  
  # Reduce distance matrix to 3 dimensions
  set.seed(123)
  mds.3D <- metaMDS(d, method="MDS", k=3, trace = 0)
  xyz <- scores(mds.3D, display="sites") # pull out the (x,y,z) coordinates
  
  # Plot
  fig.3D <- plot_ly(x=xyz[,1], y=xyz[,2], z=xyz[,3], type="scatter3d", mode="markers",
                    color=sample_data(physeq.mars)$host_disease, colors = c("blue", "red"))%>%
    layout(title = paste('MDS in 3D with', name_dist, 'distance', sep = ' '))
  
  return(fig.3D)
}

Now let’s plot!

plotDistances3D(dist.mars$UniF, "UniFrac")

## Warning in metaMDS(d, method = "MDS", k = 3, trace = 0): stress is (nearly)
## zero: you may have insufficient data

plotDistances3D(dist.mars$Ait, "Aitchison")

plotDistances3D(dist.mars$Canb, "Canberra")

plotDistances3D(dist.mars$Bray, "Bray-Curtis")

HIERARCHICAL CLUSTERING

# For heatmaps: have group color
matcol <- data.frame(phenotype = sample_data(physeq.mars)[,"host_disease"],
                     host_subtype = sample_data(physeq.mars)[,"host_subtype"],
                     collection = sample_data(physeq.mars)[,"Collection"])


# Function to get heatmap from the distances computed
plotHeatmaps <- function(dlist, fontsize){
  
  # Initialize variables
  i=1
  plist <- vector("list", 4)
  names(plist) <- names(dlist)
  
  # Loop through distances
  for(d in dlist){
    plist[[i]] <- pheatmap(as.matrix(d), 
                          clustering_distance_rows = d,
                          clustering_distance_cols = d,
                          fontsize = fontsize,
                          show_rownames = F,
                          show_colnames = F,
                          annotation_col = matcol,
                          # annotation_row = matcol,
                          annotation_colors = list(host_disease = c('Healthy' = 'blue', 'IBS' = 'red'),
                                                   host_subtype = c('HC'='black', 'IBS-M'='orange', 'IBS-C'='brown', 'IBS-D'='red'),
                                                   Collection = c("1st"='#D95F02', '2nd'='#7570B3')),
                          cluster_rows = T,
                          cluster_cols = T,
                          clustering_method = 'complete', # hc method
                          main = names(dlist)[i]) # have name of distance as title
    i <- i+1
  }
  
  return(plist)
}


# Get the heatmaps
heatmp.mars <- plotHeatmaps(dlist = dist.mars, fontsize = 8)

REPRODUCE PLOTS FROM PAPER

Figure 1 - Comparison of the Bray-Curtis beta-diversity dispersion

# get useful metadata
metadata <- sample_data(physeq.mars) %>%
  as_tibble() %>%
  select(Run, host_disease, host_subtype, host_ID)

# Build dataframe with Bray-Curtis distance between IBS-HC, IBS-IBS, HC-HC
bc.comp <- melt(as.matrix(dist.mars$Bray), varnames = c("row", "col")) %>%
  inner_join(metadata, by=c("row"="Run")) %>%
  dplyr::rename(row_disease=host_disease, row_subtype=host_subtype, row_hostID=host_ID) %>%
  inner_join(metadata, by=c("col"="Run")) %>%
  dplyr::rename(col_disease=host_disease, col_subtype=host_subtype, col_hostID=host_ID) %>%
  # Keep only comparison within subject (row_hostID == col_hostID), removing comparison with same sample (BC=0)
  filter(row_hostID == col_hostID) %>%
  filter(value != 0) %>%
  # BC(SampleA,SampleB) == BC(SampleB,SampleA), so remove duplicates
  group_by(row_hostID) %>%
  summarise_all(first) %>%
  ungroup() %>%
  select(-row_hostID, -col_hostID) %>%
  # Classify comparison as between disease subtypes
  mutate(compare_subtype=ifelse(row_subtype == "HC" & col_subtype=="HC", "Within healthy",
                                ifelse(row_subtype == "IBS-C" & col_subtype=="IBS-C", "Within IBS-C",
                                       ifelse(row_subtype == "IBS-D" & col_subtype=="IBS-D", "Within IBS-D", "Unknown")))) %>%
  mutate(compare_subtype=factor(compare_subtype, levels=c("Within IBS-C", "Within IBS-D", "Within healthy")))

# Sanity check
bc.comp %>%
  dplyr::count(row_subtype, col_subtype, compare_subtype)

# Plot fig1J
ggplot(bc.comp, aes(x=compare_subtype, y=value))+
  geom_boxplot(width=0.2, outlier.shape = NA)+
  geom_jitter(aes(color=compare_subtype), size=3, width=0.01, alpha=0.7)+
  scale_color_manual(values=c("#ff7f00", "#377eb8", "#999999"))+
  theme_cowplot()+
  theme(axis.text.x = element_text(angle=45, hjust=1),
        legend.position="none")+
  labs(x='', y='Within subject BC dissimilarity')

(J) Community variability within each group based on mean Bray Curtis Distance (HC versus IBS-C, p value = 0.02, ANOVA Tukey HSD, n = 10, 11, and 9 mucosal microbiome samples for IBS-C, IBS-D, and HC, respectively).

As a proxy of the fig1G, let’s compute the between-subject BC distance.

#___________________________________________________________________

# Build dataframe with Bray-Curtis distance between IBS-HC, IBS-IBS, HC-HC
bc.comp <- melt(as.matrix(dist.mars$Bray), varnames = c("row", "col")) %>%
  inner_join(metadata, by=c("row"="Run")) %>%
  dplyr::rename(row_disease=host_disease, row_subtype=host_subtype, row_hostID=host_ID) %>%
  inner_join(metadata, by=c("col"="Run")) %>%
  dplyr::rename(col_disease=host_disease, col_subtype=host_subtype, col_hostID=host_ID) %>%
  # Keep only comparison between different subjects (row_hostID != col_hostID)
  filter(row_hostID != col_hostID) %>%
  # Remove duplicate values (comparison(A,B) = comparison(B,A), so remove one of them)
  rowwise() %>%
  mutate(comp = paste(sort(c(row, col)), collapse="-")) %>%
  ungroup() %>%
  group_by(comp) %>%
  summarise_all(first) %>%
  ungroup() %>%
  select(-comp) %>%
  # Classify comparison as between diseases, between Healthy, or between IBS samples
  mutate(compare_disease=ifelse(row_disease != col_disease, "Between",
                        ifelse(row_disease == "Healthy", "Healthy",
                               ifelse(row_disease == "IBS", "IBS", "Unknown")))) %>%
  mutate(compare_disease=factor(compare_disease, levels=c("Between", "IBS", "Healthy"))) %>%
   # Classify comparison as between disease subtypes
  mutate(compare_subtype=ifelse(row_subtype == "HC" & col_subtype=="IBS-C", "Healthy_IBS-C",
                         ifelse(row_subtype == "IBS-C" & col_subtype=="HC", "Healthy_IBS-C",
                                ifelse(row_subtype == "HC" & col_subtype=="IBS-D", "Healthy_IBS-D",
                                ifelse(row_subtype == "IBS-D" & col_subtype=="HC", "Healthy_IBS-D",
                                       ifelse(row_subtype == "IBS-D" & col_subtype=="IBS-C", "Between IBS subtypes",
                                       ifelse(row_subtype == "IBS-C" & col_subtype=="IBS-D", "Between IBS subtypes",
                                              ifelse(row_subtype == "HC" & col_subtype=="HC", "Within healthy",
                                                     ifelse(row_subtype == "IBS-C" & col_subtype=="IBS-C", "Within IBS-C",
                                                            ifelse(row_subtype == "IBS-D" & col_subtype=="IBS-D", "Within IBS-D", "Unknown")))))))))) %>%
  mutate(compare_subtype=factor(compare_subtype, levels=c("Within IBS-C", "Healthy_IBS-C", "Within healthy", "Healthy_IBS-D", "Within IBS-D", "Between IBS subtypes")))

# Sanity check
bc.comp %>%
  dplyr::count(row_disease, col_disease, compare_disease)

bc.comp %>%
  dplyr::count(row_subtype, col_subtype, compare_subtype)

# Attempt at reproducing fig1G
ggplot(bc.comp, aes(x=compare_subtype, y=value))+
  geom_boxplot(width=0.2, outlier.shape = NA)+
  geom_jitter(aes(color=compare_subtype), size=2, width=0.01, alpha=0.5)+
  scale_color_manual(values=c("#ff7f00", "#a1d76a", "#999999", "#9ecae1", "#2171b5", "#807dba"))+
  theme_cowplot()+
  theme(axis.text.x = element_text(angle=45, hjust=1),
        legend.position="none")+
  labs(x='', y='Bray-Curtis dissimilarity')

(G) Community variability determined by the mean within-subject Bray Curtis distance (within-IBS-D versus within-IBS-C, p = < 0.005, ANOVA Tukey, n = 22, 46, 24, 53, and 29, Bray-Curtis distances between stool samples of the same subject for within-IBS-C, Healthy versus IBS-C, within-Healthy, Healthy versus IBS-D, and within-IBS-D, respectively).

Just as personal interest, let’s look at the disease level (IBS-HC-IBSvsHC):

# Plot within IBS, within Healthy, and between IBS-Healthy BC dissimilarity
ggplot(bc.comp, aes(x=compare_disease, y=value))+
  geom_violin()+
  geom_boxplot(width=0.1, outlier.shape = NA)+
  geom_jitter(size=0.1, width=0.01)+
  theme_cowplot()+
  labs(x='', y='Bray-Curtis dissimilarity')